Nanoscale
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
PAPER
Cite this: Nanoscale, 2015, 7, 8122
View Journal | View Issue
Synthesis-atomic structure-properties
relationships in metallic nanoparticles by
total scattering experiments and 3D computer
simulations: case of Pt–Ru nanoalloy catalysts†
Binay Prasai,a Yang Ren,b Shiyao Shan,c Yinguang Zhao,c Hannah Cronk,c Jin Luo,c
Chuan-Jian Zhongc and Valeri Petkov*a
An approach to determining the 3D atomic structure of metallic nanoparticles (NPs) in ﬁne detail and
using the unique knowledge obtained for rationalizing their synthesis and properties targeted for optimization is described and exempliﬁed on Pt–Ru alloy NPs of importance to the development of devices for
clean energy conversion such as fuel cells. In particular, PtxRu100−x alloy NPs, where x = 31, 49 and 75, are
synthesized by wet chemistry and activated catalytically by a post-synthesis treatment involving heating
under controlled N2–H2 atmosphere. So-activated NPs are evaluated as catalysts for gas-phase CO oxidation and ethanol electro-oxidation reactions taking place in fuel cells. Both as-synthesized and activated NPs are characterized structurally by total scattering experiments involving high-energy
synchrotron X-ray diﬀraction coupled to atomic pair distribution functions (PDFs) analysis. 3D structure
models both for as-synthesized and activated NPs are built by molecular dynamics simulations based on
the archetypal for current theoretical modelling Sutton–Chen method. Models are reﬁned against the
experimental PDF data by reverse Monte Carlo simulations and analysed in terms of prime structural
characteristics such as metal-to-metal bond lengths, bond angles and ﬁrst coordination numbers for Pt
and Ru atoms. Analysis indicates that, though of a similar type, the atomic structure of as-synthesized and
respective activated NPs diﬀer in several details of importance to NP catalytic properties. Structural
characteristics of activated NPs and data for their catalytic activity are compared side by side and strong
evidence found that electronic eﬀects, indicated by signiﬁcant changes in Pt–Pt and Ru–Ru metal bond
lengths at NP surface, and practically unrecognized so far atomic ensemble eﬀects, indicated by distinct
stacking of atomic layers near NP surface and prevalence of particular conﬁgurations of Pt and Ru atoms
in these layers, contribute to the observed enhancement of the catalytic activity of PtxRu100−x alloy NPs at
Received 3rd February 2015,
Accepted 3rd April 2015
x ∼ 50. Implications of so-established relationships between the atomic structure and catalytic activity of
Pt–Ru alloy NPs on eﬀorts aimed at improving further the latter by tuning-up the former are discussed
DOI: 10.1039/c5nr00800j
and the usefulness of detailed NP structure studies to advancing science and technology of metallic
www.rsc.org/nanoscale
NPs – exempliﬁed.
Introduction
With current science and technology moving rapidly into
smaller scales, metallic nanoparticles (NPs) are synthesized in
a
Department of Physics, Central Michigan University, Mt. Pleasant,
Michigan 48858, USA. E-mail: [email protected]
b
X-ray Science Division, Advanced Photon Source, Argonne National Laboratory,
Argonne, Illinois 60439, USA
c
Department of Chemistry, State University of New York at Binghamton,
New York 13902, USA
† Electronic supplementary information (ESI) available: XRD patterns, TEM and
3D structure modelling methodology. See DOI: 10.1039/c5nr00800j
8122 | Nanoscale, 2015, 7, 8122–8134
increasing numbers and explored for various useful applications ranging from catalysis1–3 and photonics4,5 to magnetic
storage media6–8 and drug delivery.9–11 Catalytic applications
are of particular importance since, at present, virtually all
transportation fuels and most chemicals are produced by
technologies based on catalysis.12 Among others, a few nanometre in size Noble metals-based alloy particles have shown a
great promise as catalysts for several technologically important
reactions because not only the particles have a large surface to
volume ratio but also allow achieving higher catalytic activity
and better selectivity by exploiting the synergy of metallic
species alloyed. Advancing science and technology of metallic
NPs, including metallic nanoalloy catalysts, however, face a
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
major hurdle concerning transforming the current, largely
trial-and-error route to synthesizing metallic NPs into synthesizing and optimizing metallic NPs for practical applications
by rational design. This is not a trivial task since the outcome
of synthesis and optimization of metallic NPs, typically done
relevant by post-synthesis treatment, depend on a number of
hard to assess kinetic and thermodynamic factors. Among
others, an important step toward overcoming the hurdle would
be the implementation and validation of an approach to determining the outcome of synthesis and optimization of metallic
NPs at atomic level, that is the atomic-level structure of actual
metallic NPs produced, and then using the unique knowledge
obtained for rationalizing NP synthesis and properties targeted
for optimization.
Indeed several experimental techniques such as coherent
diﬀraction,13 high resolution Transmission Electron
Microscopy (HR-TEM),14 Extended X-ray Absorption Fine Structure (EXAFS) spectroscopy,15 and others, have proven useful in
structure studies of metallic NPs, including nanoalloy catalysts. However, due to their character, these techniques have
not been able to determine the 3D atomic structure in fine
detail but reveal some structural features of metallic NPs only.
Recently total scattering coupled to atomic pair distribution
functions (PDFs) analysis have emerged as a very eﬃcient technique for structural characterization of materials confined to
nanoscale dimensions, including metallic NPs.16 Here we
augment the technique with 3D computer simulations and
apply it to Pt–Ru nanoalloy catalysts. Among other families of
metallic alloy NPs explored for catalytic applications, Pt–Ru
one was chosen because of its importance for the development
of currently highly pursued devices for clean energy conversion
such as fuel cells, in particular for the family’s remarkable
ability to speed up gas-phase CO oxidation and ethanol
electro-oxidation reactions taking place in fuel cells. The
former entails the application of Pt–Ru alloy NPs as catalysts
for the hydrogen oxidation reaction (HOR) taking place at the
anode of Proton Exchange Membrane Fuel Cells (PEMFCs).
The latter entails the application of Pt–Ru alloy NPs as catalysts in liquid fuel cells. Note, nanoalloy catalysts for HOR in
PEMFCs are very much needed since, in the current state of
technology, inexpensive hydrogen gas comes mostly from
steam reforming of hydrocarbons. “Reformate” hydrogen,
however, contains some impurities such as CO poisoning
monometallic Pt NPs typically used as PEMFC anode catalysts
leading to a severe degradation of PEMFCs performance. Supported Pt–Ru bulk and surface alloy NPs have emerged as a
CO-tolerant and so very eﬃcient PEMFC anode catalysts thus
generating considerable research interest recently.17–21 Pt–Ru
alloy NPs have also shown good promise as electro-catalysts for
methanol oxidation in direct methanol fuel cells (DMFCs)
wherein the oxidation of CO species resulted from methanol
decomposition needs to be promoted.22–25 Here we pay particular attention to the electro-catalytic properties of Pt–Ru
alloy NPs not in methanol but in direct ethanol fuel cells
(DEFCs). Those are very attractive since ethanol not only
has higher energy density (8.0 kW h kg−1 vs. 6.1 kW h kg−1)
This journal is © The Royal Society of Chemistry 2015
Paper
than methanol but also is bio-renewable and not toxic. A key
challenge to the commercial viability of DEFCs is the development of catalysts that can alleviate the formation of undesired
by-products from ethanol decomposition such as acetic acid
and acetaldehyde by eﬀective cleavage of C–C bonds and accelerate the oxidation of CO species poisoning the reaction. Pt–
Ru alloy NPs have shown very good promise in this regard as
well.26,27 However, the numerous studies on Pt–Ru alloy NPs
carried out so far have focused mostly on their catalytic properties and so provided limited knowledge about NP atomic
structure. Here we concentrate on determining the atomic
structure of Pt–Ru alloy NPs in fine detail, establishing the
relationships between their atomic structure and catalytic properties, and demonstrating how these relationships can
provide a feedback loop for streamlining synthesis and so optimizing performance of Pt–Ru alloy NPs in fuel cells related
applications. For the purpose:
(i) Synthesized PtxRu100−x alloy NPs (x = 31, 49 and 75) by
wet chemistry and activated them catalytically by a controlled
post-synthesis treatment involving heating in N2–H2
atmosphere.
(ii) Determined the morphology, i.e. size and shape, of both
as-synthesized and post-synthesis treated Pt–Ru alloy NPs by
TEM.
(iii) Determined the chemical pattern of Pt and Ru species
across NPs by High-Angle Annular Dark-Field (HAADF) Scanning TEM (STEM) experiments.
(iv) Characterized structurally both as-synthesized and postsynthesis treated Pt–Ru NPs by total scattering experiments
involving high-energy synchrotron X-ray diﬀraction (XRD) and
atomic PDFs analysis.16,28–30
(v) Built 3D structure models for both as-synthesized and
post-synthesis treated NPs by Molecular Dynamics (MD) simulations based on the archetypal for current metal and alloy
structure theory Sutton–Chen (SC) method.31,32 Simulations
required developing SC potential for Ru since, contrary to the
case of face-centered-cubic (fcc) metals such as Pt, SC potentials for hexagonal-close-packed (hcp) metals such as Ru are
not readily available.
(vi) Refined MD models by reverse Monte Carlo (RMC)
simulations guided by the experimental atomic PDFs.
(vii) Analysed RMC refined models in terms of structural
characteristics important for catalysis such as NP phase state,
distribution of first atomic neighbour distances, i.e. metal-tometal bond lengths, first atomic coordination numbers and
bond angles paying particular attention to Pt and Ru atoms at
NP surface since it is the NP surface where catalytic reactions
take place.
(viii) Measured the catalytic activity of post-synthesis
treated, i.e. activated for catalytic applications, Pt–Ru alloy
NPs for gas-phase oxidation of CO and electro-oxidation of
ethanol.
(ix) Correlated data for catalytic activity with structural
characteristics of respective NPs and established which of the
latter are most likely to contribute to the observed enhancement of the former at x ∼ 50.
Nanoscale, 2015, 7, 8122–8134 | 8123
View Article Online
Paper
(x) Based on so-established relationships between the
atomic structure and catalytic activity of Pt–Ru alloy NPs
suggested routes to improving further the latter by tuning up
the former.
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Experimental
PtxRu100−x NPs (x = 31, 49 and 75) synthesis and post-synthesis
treatment
Pt–Ru alloy NPs were synthesized by a protocol fully described
in ref. 33. In brief, platinum(II) acetylacetonate and ruthenium(II)
acetylacetonate were mixed in the desired molar ratio in
octyl ether solution under N2 atmosphere. Oleic acid and oleylamine were added to the solution as NP capping agents and
1,2-hexadecanediol – as a reducing agent. The solution was
purged with N2, heated to 220 °C and kept at that temperature
for 30 min. After the solution cooled back to room temperature, the resulting Pt–Ru NPs were precipitated out by adding
ethanol followed by centrifugation. Precipitated NPs were
redispersed in hexane and mixed with fine carbon powder
(XC-72) followed by sonication and overnight stirring. The
resulting carbon supported Pt–Ru NPs, hereafter referred to as
“as-synthesized”, were collected and dried under N2 atmosphere. Then, in line with common practices, carbon supported
Pt–Ru NPs were activated for catalytic applications. The activation involved a two-step thermal treatment under controlled
gas atmosphere.34 In particular, at first NPs were heated to
260 °C under N2 atmosphere aiming at removing the organic
molecules capping their surface. Next, NPs were further heated
to 400 °C under 15 vol% H2 atmosphere and kept at that temperature for 2 h. So post-synthesis treated NPs hereafter are
referred to as “fully activated catalytically”. The loading of Pt–
Ru NPs on the carbon support was determined by thermogravimetric analysis (TGA) performed on a Perkin-Elmer Pyris
1-TGA instrument, and found to be within 15–20% by weight.
PtxRu100−x NPs (x = 31, 49 and 75) size, shape, chemical
composition and chemical pattern determination
The size and shape of as-synthesized and fully activated Pt–Ru
NPs were determined by TEM. For the measurements batches
of NPs were diluted in hexane and drop cast onto carboncoated copper grids followed by solvent evaporation in air at
room temperature. The measurements were done on
JEM-2200FS microscope operated at 200 kV. The microscope
was fitted with an ultra-high-resolution (UHR) pole piece with
a point resolution of 0.19 nm. Exemplary TEM and highresolution (HR)-TEM images of as-synthesized and fully activated PtxRu100−x alloy NPs (x = 31, 49 and 75) are shown in
Fig. S1 and S2,† respectively. As can be seen in the figures (i)
as-synthesized NPs are approximately 4.3(±0.6) nm in size and
spherical in shape, (ii) fully activated NPs largely retain the
spherical shape of as-synthesized ones and slightly increase in
size to approximately 4.6(±0.7) nm and (iii) all NPs exhibit
good overall crystallinity and, as usual for metallic NPs, some
surface structural disorder.
8124 | Nanoscale, 2015, 7, 8122–8134
Nanoscale
The chemical composition of as-synthesized and fully activated PtxRu100−x NPs (x = 31, 49, 75) was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES)
carried out on a batch of samples for each x. Experiments were
done on a Perkin Elmer 2000 DV ICP-OES instrument using a
cross flow nebulizer with the following parameters: plasma
18.0 L Ar(g) min−1; auxiliary 0.3 L Ar(g) min−1; nebulizer 0.73 L
Ar(g) min−1; power 1500 W; peristaltic pump rate 1.40 mL
min−1. Samples were dissolved in aqua regia, and then diluted
to concentrations in the range of 1 to 50 ppm. Calibration
curves were made from dissolved standards in the same acid
matrix as the unknowns. Standards and the unknowns were
analysed 10 times each resulting in <3% error in the reported
chemical composition. Within the error limits no appreciable
diﬀerence between the chemical composition of as-synthesized
and respective fully activated Pt–Ru NPs was observed. Note,
similar results, i.e. that a thermo-chemical activation of Noble
metals-based NPs for catalytic applications conducted with
due care does not necessarily induce significant changes in NP
overall chemical composition, size and shape, have been
reported by Dash et al.35
The chemical pattern of fully activated PtxRu100−x NPs (x =
31, 49, 75) was determined by High-Angle Annular Dark-Field
(HAADF) Scanning TEM (STEM) experiments. Experiments
were done on a JEOL JEM 2100F instrument equipped with a
CEOS hexapole probe. The instrument was operated at 200 keV
in STEM mode. The lens settings combined with the corrector
tuning gave a spatial resolution of ∼90 pm. Exemplary HAADFSTEM images are shown in Fig. S3.† Images indicate that fully
activated Pt49Ru51 and Pt75Ru25 NPs are largely monophase
alloys while Pt31Ru69 NPs are partially segregated alloys.
Atomic PDFs analysis and 3D structure modelling described
below confirmed the results of HAADF-STEM experiments.
Total scattering experiments
Carbon supported as-synthesized and respective fully activated
PtxRu100−x alloy NPs (x = 31, 49 and 75) were subjected to highenergy synchrotron X-ray diﬀraction (XRD) experiments at the
11-ID-C beamline of the Advanced Photon Source, Argonne.
X-rays of energy 115 keV (λ = 0.1080 Å) were used. Diﬀraction
data was collected on a large area detector up to wave vectors
of 25 Å−1 allowing resolving structural features of NPs in fine
detail. The experimental set-up was calibrated with high-purity
powder Si standard. During the measurements samples were
sealed in thin-walled glass capillaries. An empty glass capillary
and carbon powder alone were measured separately. Experimental synchrotron high-energy XRD patterns corrected for
sample absorption, empty glass capillary, carbon support and
other background-type (air etc.) scattering are shown in
Fig. S4.† Patterns exhibit a few broad, strongly overlapping
peaks at low diﬀraction (Bragg) angles and almost no distinct
peaks at high diﬀraction angles, i.e. are rather diﬀuse in
nature. This rendered the well-established, sharp Bragg peaksbased procedures for determining the atomic-scale structure
of bulk metals and alloys inapplicable to Pt–Ru alloy NPs
studied here. Therefore, the diﬀuse XRD patterns were reduced
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
Paper
Fig. 1 Experimental (symbols) and model (lines) PDFs for as-synthesized (a) and post-synthesis treated (b) PtxRu100−x alloy NPs (x = 31, 49 and 75).
Model PDFs based on a hcp-type crystal structure are given in blue while those based on a fcc-type crystal structure – in red. Values of “hcp- and
fcc-lattice parameters”, derived by adjusting structure models against the respective experimental PDFs, are given by each data set. Model’s quality
indicators, RPDF
wp , explained in ESI,† are shown by each data set in color pertaining to the respective structure model. Experimental and model atomic
PDFs for ∼7 nm in size pure Pt particles and ∼3.5 nm in size pure Ru particles, obtained by independent studies, are also shown in (a) to facilitate
diﬀerentiating between the structural features of hcp(Ru) and fcc(Pt)-type models.
to the so-called atomic PDFs G(r) following well-established
procedures.36 The approach is advantageous28–30 since, contrary to the respective diﬀuse XRD patterns, atomic PDFs show
several distinct peaks allowing convenient testing and refinement of atomic-structure models for NPs. Atomic PDFs G(r)
derived from the XRD patterns of Fig. S4† are shown in Fig. 1.
Note, XRD patterns of Fig. S4† and so their Fourier counterparts, the atomic PDFs of Fig. 1, reflect ensemble averaged
structural features of all Pt–Ru NPs sampled by the X-ray beam
in a way traditional powder XRD patterns reflect ensemble
averaged structural features of all polycrystallites sampled by
the X-ray beam in those experiments. Using NP ensemble averaged structural features to understand and explain NP ensemble averaged properties (e.g. catalytic, magnetic, optical etc.)
puts NP atomic structure–properties exploration on the same
footing. Also, note, by definition, atomic PDFs G(r) peak at distances separating all pairs of atoms within the material
studied. Accordingly, the experimental PDFs of Fig. 1 peak at
distances separating Pt–Pt, Pt–Ru and Ru–Ru pairs of atoms,
immediate and all farther neighbours, within Pt–Ru NPs
studied here. Since atoms at/close to opposite sides of NPs are
separated most, the atomic PDFs of Fig. 1 may show distinct
peaks up to distances close to the average size (diameter) of
the respective NPs. As can be seen in Fig. 1, however, peaks in
the experimental PDFs decay to zero at distances (∼3.5 nm)
shorter than the physical size of respective NPs, which is
∼4.5 nm as determined by TEM. The observed fast decay of
This journal is © The Royal Society of Chemistry 2015
the experimental PDFs indicates that surface atoms in Pt–Ru
alloy NPs suﬀer non-negligible positional disorder. Such is
common to metallic NPs due to NP finite size and surface
relaxation eﬀects.13–15,28–30
Catalytic activity measurement
The catalytic activity of fully activated PtxRu100−x alloy NPs (x =
31, 49 and 75) for gas-phase oxidation of CO was measured
under CO (0.5 vol% balanced by N2) + O2 (10 vol% balanced by
N2) atmosphere using a custom-built system including a temperature-controlled reactor, a gas flow/mixing/injection controller, an on-line gas chromatograph (Shimadzu GC 8A) equipped
with 5A molecular sieve, Porapak Q packed columns and a
thermal conductivity detector. Carbon supported NPs were
loaded in a quartz microreactor tubing with an inner diameter
of 4 mm and the tubing was wrapped with quartz wool. The
feeding gas (0.5 vol% CO + 10 vol% O2 balanced by N2) was
injected continuously through the fixed NPs’ bed (∼6 mm in
length) at a flow rate of 20 mL min−1. The residence time was
about 0.2 seconds. Gas hourly space velocity (GHSV) in the
system was about 16 000 h−1. The amount of highly reactive
and toxic CO oxidized to largely inert and harmless CO2 was
determined by analysing the composition of tail gas eﬀusing
from the quartz microreactor using the on-line (Shimadzu GC
8A) gas chromatograph.
The catalytic activity of fully activated PtxRu100−x alloy NPs
(x = 31, 49 and 75) for electro-oxidation of ethanol was
Nanoscale, 2015, 7, 8122–8134 | 8125
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Paper
measured following a widely adopted procedure described in
ref. 37. In brief, for each x, a suspension of the carbon supported NPs was prepared by dissolving 10 mg of NPs in 10 ml
of water containing 5% vol. of Nafion polymer. The suspension
was ultrasonicated for 10 min until it turned into a dark ink.
10 μL of the ink were uniformly distributed over the surface of
a polished, glassy carbon disk with a surface area of
0.196 cm2. The disc, as coated with a thin film of the ink, was
dried overnight at room temperature, inserted into an electrochemical cell and used as cell’s working electrode. The cell
included two more electrodes, one known as a reference (Ag/
AgCl, saturated KCl) electrode and the other – as a counter (Pt)
electrode. Those were positioned in separate compartment of
the cell. Electrolyte was made by dissolving KOH pellets
(optimal grade) in Milli-Q water as to achieve a concentration
of 0.5 M KOH. Ethanol was added until its concentration in
the electrolyte also reached 0.5 M level. The electrolyte was
deaerated with high purity N2 before conducting cyclic voltammetry (CV) sweeps at room temperature. Current generated by
the cell was measured while the potential of the working electrode was varied between −0.9 and +0.5 V. Note, in CV
measurements, typically, “forward current” pertains to the
current generated while the potential of working electrode
goes from negative (−0.9 V) to positive (+0.5 V). Accordingly,
“backward current” pertains to the current generated while the
potential of working electrode goes from positive (+0.5 V) to
negative (−0.9 V). Data for the electrocatalytic activity of fully
activated Pt–Ru alloy NPs presented here reflect the peak
values of “forward” current as normalized against the amount
of Pt loaded in the catalyst on the working electrode.
Results and discussion
When alloyed in bulk, fcc Pt and hcp Ru straddle a stability
limit in the binary phase diagram even when annealed at
temperature above 1000 °C.38 In particular, solids containing
more than 50 at% of Pt tend to crystallize in a fcc-type structure. On the other hand, solids containing less than 20 at% of
Pt tend to crystallize in a hcp-type structure. Solids containing
from 20 to 50 at% Pt tend to segregate into coexisting hcp(Rurich) and fcc(Pt-rich) type phases. PtxRu100−x alloy NPs (x = 31,
49 and 75) studied here were synthesized and post-synthesis
treated at temperature much lower than 1000 °C. Therefore,
the type of their atomic-scale structure could not be inferred
for sure from that of Pt–Ru solids of similar chemical composition. Besides, studies have shown that metals and alloys confined to nanoscale dimensions do not necessarily adopt the
structure type of their bulk counterparts.39–41 The structure
type of Pt–Ru alloy NPs studied here was determined by
approaching the respective experimental atomic PDFs with
purely empirical models constrained to the fcc- and hcp-type
perfectly 3D periodic, infinite lattices used to describe the
crystal structure of bulk Pt and Ru, respectively. The δ-function
like peaks in the atomic PDFs derived from the model lattices
were broadened by convolution with gaussian functions as to
8126 | Nanoscale, 2015, 7, 8122–8134
Nanoscale
mimic the dynamic (i.e. thermal) and static structural disorder
usually present in metallic NPs. At the same time model PDFs
were multiplied by an artificial, rapidly decaying with real
space distances function as to mimic the finite size (∼4.5 nm)
of Pt–Ru NPs. Finally, the unit cell parameters of model fccand hcp-lattices were refined so that the model-derived atomic
PDFs approached the experimental ones as close as possible.
Here it may be noted that “lattice parameters” of metallic NPs
are not as well defined as a physical quantity as those for their
bulk counterparts. The reason is that “lattice parameters”
imply the presence of perfectly 3D periodic, infinite lattices
while metallic NPs are finite and not necessarily perfectly 3D
periodic at atomic level. Nevertheless, total scattering/PDF
data-derived “lattice parameters” of metallic NPs are useful
merely because they reflect the set of interatomic distances,
including metal-to-metal bond lengths, characteristic for the
actual NPs under study and, hence, may be used as a trustworthy “global indicator” of those distances. Modeling was
done with the help of program PDFgui.42 Results from the
modeling are shown in Fig. 1.
As can be seen in the figure, models featuring a fcc-type
crystal structure reproduce very well (RPDF
∼ 19–21%) the
wp
experimental PDFs for as-synthesized and respective fully activated Pt75Ru25 NPs indicating that atoms in these NPs are fcctype ordered. Models featuring a fcc-type crystal structure
reproduce the experimental PDFs for as-synthesized and
respective fully activated Pt49Ru51 NPs at an acceptable level
(RPDF
wp ∼ 23–25%) indicating that atoms in these NPs are largely
but not entirely fcc-like ordered. Some features in the experimental PDFs for as-synthesized and respective fully activated
Pt31Ru69 NPs are reproduced well by a fcc-type model while
others – by a hcp-type one. However, neither a fcc- (RPDF
wp ∼
34–40%) nor a hcp-type (RPDF
wp ∼ 40–46%) model alone reproduce the experimental PDF data well indicating that atoms in
Pt31Ru69 NPs studied here are, very likely, segregated into
domains each with a distinct fcc- or hcp-type structure.
Inspection of data in Fig. 1 also reveals that the characteristic interatomic distances both in as-synthesized and fully
activated PtxRu100−x alloy NPs (x = 31, 49 and 75), as reflected
by the respective PDF-derived “lattice parameters”, are significantly diﬀerent from those in bulk fcc Pt (a = 3.924 Å) and hcp
Ru (a = 2.704 Å and c = 4.281 Å). Furthermore, it reveals that
the post-synthesis treatment of NPs (i) diminishes atomic positional disorder in Pt75Ru25 NPs (see the diﬀerence in the
sharpness of the peaks in the respective PDFs), (ii) does not
aﬀect atomic positional disorder in Pt49Ru51 NPs substantially
(see the similar sharpness of the peaks in the respective PDFs)
and (iii) increases atomic positional disorder in Pt31Ru69 NPs
(see the faster decay of the atomic PDF for post-synthesis
treated sample to zero). Indeed (i) to (iii) are easier to acknowledge by inspecting data in Fig. 4 introduced later on. In
general, the empirical modeling indicated that PtxRu100−x
alloy NPs (x = 31, 49 and 75) studied here adopt a fcc- and/or
hcp-like atomic structure but exhibit a unique degree of
atomic positional disorder and/or partial segregation for any
given x.
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
Accounting for the findings of simplistic empirical modeling realistic 3D models both for as-synthesized and post-synthesis treated Pt–Ru alloy NPs were built by MD simulations
based on the archetypal for current metal and alloy atomic
structure theory Sutton–Chen (SC) method. Among other
theoretical methods currently in use, SC one was chosen
because of its simplicity and proven success.43,44 Method’s
basics are given as ESI.† SC parameters for Pt species adopting
a fcc-type structure in bulk were taken from literature
sources.31,32 Parameters are listed in Table S1.† SC parameters
for Ru species adopting a hcp-type structure in bulk were
derived by a well-established procedure described in ESI.†
Parameters are listed in Table S2.† MD models featured
NPs with the chemical composition, chemical pattern
(see Fig. S3†), size (∼4.5 nm) and shape (spherical; see
Fig. S1 and S2†) of actual Pt–Ru alloy NPs studied here.
Details of MD simulations are given as ESI.† Resulted 3D
models are shown in Fig. S5.† Atomic PDFs derived from the
models are compared with the respective experimental PDFs
in Fig. 2.
As can be seen in Fig. 2 MD built structure models produce
atomic PDFs that capture the essence of experimental ones but
fail in reproducing them in fine detail. In particular, MD
models fail in reproducing the position, broadening and intensity of several peaks in the experimental PDFs, i.e. the radii,
degree of atomic positional disorder and atomic population of
several coordination spheres in Pt–Ru alloy NPs. The inability
of MD models to reproduce accurately the radii of atomic
coordination spheres in PtxRu100−x alloy NPs (x = 31, 49 and
75) may be expected since SC method31,32,45 employed by MD
is, like most theoretical methods currently in use, validated
Paper
against experimental data for the lattice parameters of respective solids (see Tables S1 and S2†). Results of empirical modeling carried out here, however, showed clearly that NP “lattice
parameters”, i.e. the characteristic interatomic distances in Pt–
Ru alloy NPs, are diﬀerent from these of the respective solids.
Indeed this is a rather common observation since interatomic
distances in metallic NPs can get “compressed” or “stretched
out” considerably and non uniformly as a result of the inherently increased surface tension of NPs13 and/or NP surface
environment eﬀects such as, for example, exposure to reactive
(e.g. H2) gases.46 Such an eﬀect is atypical for the respective
solids. Besides, charge transfer and optimization of atomiclevel stresses in metallic alloy NPs can result in a considerable
change in the radii/size of metallic species involved which,
again, may not be the case with the respective solids.16,47 The
inability of MD models to reproduce accurately the degree of
local structural disorder in PtxRu100−x alloy NPs (x = 31, 49 and
75), in particular its inherent enhancement in coordination
spheres involving atoms close to/at the extended NP surface,
may also be expected. It is due to the fact that MD treats all
atoms in structure models alike whereas atoms inside and at
the surface of real-world metallic NPs are not necessarily alike
in respect to their immediate atomic neighborhood, and
others. The inability of MD models to reproduce accurately the
distribution of Pt and Ru atoms across NPs is not a big surprise either. The reason is that, typically, MD models feature
atomic configurations quenched from some high temperature
initial state while Pt–Ru alloy NPs studied here are synthesized
in solution kept at room temperature and only then subjected
to a treatment at elevated but not very high temperature.
Nevertheless, as discussed in ref. 16, MD simulations are
Fig. 2 Experimental (symbols) and model (lines in blue) PDFs for as-synthesized (a) and post-synthesis treated (b) PtxRu100−x alloy NPs (x = 31, 49
and 75). Model PDFs are derived from 3D atomic conﬁgurations (see Fig. S5†) generated by MD simulations based on SC method as described in the
text. Model’s quality indicators, RPDF
wp , explained in ESI,† are shown by each data set.
This journal is © The Royal Society of Chemistry 2015
Nanoscale, 2015, 7, 8122–8134 | 8127
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Paper
Fig. 3 RMC reﬁned 3D structure models for as-synthesized (ﬁrst row)
and post-synthesis treated (second row) PtxRu100−x alloy NPs (x = 31, left
column); (x = 49, middle column) and (x = 75; right column). Models
include approximately 3500 Pt and Ru atoms in due proportions. Ru
atoms are in green and Pt – in gray. Note while Pt and Ru atoms both in
as-synthesized and post-synthesis treated Pt49Ru51 and Pt75Ru25 NPs are
well mixed together those in Pt31Ru69 NPs segregate ( partially) into
intertwined Pt-rich and Ru-rich domains. Domains in as-synthesized
NPs seem to evolve somewhat when NPs are subjected to a post-synthesis treatment. The evolution is toward an “onion-like” chemical
pattern wherein the “onion” core and surface get enriched in Pt atoms
while the intermediate section of “onion” – in Ru atoms.
useful in providing plausible structure models for metallic NPs
that can be refined further, when necessary.
To account for the combined eﬀects of NP finite size,
inherent and/or post-synthesis treatment induced NP “compression/stretching out”, likely optimization of atomic-level
Nanoscale
stresses and charge transfer taking place upon alloying of Pt
and Ru species at the nanoscale, MD models were refined
further by RMC simulations. In the simulations positions of
atoms in MD models of Fig. S5† were adjusted, including
switching the positions of nearby Pt and Ru atoms, as that the
diﬀerence between RMC model derived and respective experimental PDF data is reduced as much as possible. At the same
time model’s energy was minimized, i.e. model’s stability
maximized, to the fullest possible extent using reliable pairwise (Lennard-Jones type) potentials. More details of RMC
simulations are given as ESI.† RMC refined models for
PtxRu100−x alloy NPs (x = 31, 49 and 75) are shown in Fig. 3.
Atomic PDFs derived from the models are compared with the
respective experimental PDFs in Fig. 4.
As can be seen in Fig. 4 atomic PDFs derived from RMC
refined 3D structure models for Pt–Ru alloy NPs reproduce
respective experimental atomic PDFs in very good detail (see
the reported RPDF
factors), including the region of longer
wp
interatomic distances/higher r-values which, as discussed
above, is sensitive to structural features of NP surface. Precise
accounting for the latter is important for catalysis since it is
the NP surface where catalytic reactions take place.
With 3D models of realistic size, shape, chemical composition, chemical species ordering pattern, type of atomic ordering and atomic coordination spheres, including coordination
radii and numbers, tuned up against relevant experimental
PDF data at hand all important structural characteristics of
actual metallic NPs studied can be derived. Characteristics can
be used for (i) revealing the outcome of particular synthesis
and post-synthesis treatment protocol employed, (ii) considering NP properties of interest on a sound structural basis and
(iii) establishing NP atomic structure–properties relationships.
Fig. 4 Experimental (symbols) and model (lines in blue) PDFs for as-synthesized (a) and post-synthesis treated (b) PtxRu100−x alloy NPs (x = 31, 49
and 75). Model PDFs are based on RMC reﬁned atomic conﬁgurations shown in Fig. 3. Model’s quality indicators, RPDF
wp , explained in ESI,† are shown
by each data set.
8128 | Nanoscale, 2015, 7, 8122–8134
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
Here we concentrate on structural characteristics relevant to
catalytic properties of metallic NPs48–62 such as metal-tometal bond lengths, bond angles and coordination numbers
(CNs) paying particular attention to metallic species at NP
surface. Note, bond angles are very sensitive to the type of
local atomic ordering while metal-to-metal bond lengths and
coordination numbers – to the strength of interactions
between metallic species and how they configure in space,
respectively. Metal-to-metal bond lengths inside and at the
surface of as-synthesized and post-synthesis treated Pt–Ru
alloy NPs, as derived from the respective RMC refined models,
are presented in Tables S5 and S6,† respectively. First CNs
inside and at the surface of NPs, as derived from the same
models, are presented in Tables S7 and S8,† respectively. Distribution of Ru–Ru–Ru and Pt–Pt–Pt bond angles across as-synthesized and post-synthesis treated Pt–Ru alloy NPs are shown
in Fig. S6 and S7,† respectively. Distribution of bond angles
between all metallic species at the surface of post-synthesis
treated NPs are shown in Fig. S8.† Data in Fig. S6–S8† also are
derived from the models shown in Fig. 3, second row. Note,
Pt–Pt–Pt and Ru–Ru–Ru bond angle distributions shown in
Fig. S6 and S7† involve the immediate Pt neighbors of each Pt
and the immediate Ru neighbors of each Ru species, respectively, across Pt–Ru alloy NPs., i.e. are sensitive to the type of
metallic species involved. On the other hand, distributions of
bond angles shown in Fig. S8† disregard the type of metallic
species involved, i.e. are sensitive to the way metallic species
pack altogether. Furthermore, they involve atoms forming the
top three layers alone in post-synthesis treated Pt–Ru alloy
NPs. Note, for fcc/hcp-type metallic particles ∼4.5 nm in size
the top three layers include ∼35% of all atoms within the NPs.
Being a volume sensitive technique total scattering coupled to
atomic PDFs analysis is fully capable of revealing details in the
atomic ordering in such large segments of NPs studied.
Data in Tables S7 and S8† indicate that the synthesis and
post-synthesis treatment protocol employed here yield PtxRu100−x
NPs (x = 31, 49 and 75) wherein Pt and Ru atoms are packed
closely both inside NPs (first CNs ∼11.7) and at NPs surface
(first CNs ∼7.3). Data in Fig. S6 and S7† indicate that, locally,
Pt atoms alone both in as-synthesized and post-synthesis
treated Pt–Ru NPs are packed in a fcc-type manner maintaining the type of atomic ordering in bulk Pt. Ru atoms alone in
as-synthesized and post-synthesis treated Ru-poor, i.e. Pt75Ru25
NPs also are packed in a fcc-type manner evidently adopting
the type of packing of majority Pt atoms. On the other hand,
Ru atoms alone in as-synthesized and post-synthesis treated
Ru-rich, i.e. Pt31Ru69 NPs, and, to a certain extent, in Pt49Ru51
NPs appear packed in a hcp-type manner largely maintaining
the type of atomic ordering in bulk Ru. Results are fully in line
with the empirical modeling described above indicating that
as-synthesized and post-synthesis treated Pt75Ru25 NPs clearly
exhibit a fcc-type structure while as-synthesized and post-synthesis treated Pt31Ru69 NPs are partially segregated into
domains each with a distinct fcc- or hcp-type structure. According the more sophisticated MD and RMC modeling these
domains are rich in Pt and Ru species, respectively. Obviously,
This journal is © The Royal Society of Chemistry 2015
Paper
both the close packed, metallic nature and the fcc and/or
hcp-type ordering, of Pt and Ru species in PtxRu100−x NPs
(x = 31, 49 and 75) do not change significantly upon NP activation for catalytic applications by a thermo-chemical treatment of the type employed here. However, as data in Tables S5
and S6† indicate, finer details in NP atomic structure, such as
first atomic neighbor distances, i.e. metal-to-metal bond
lengths, change. In particular, Pt–Pt and Ru–Ru metal bond
lengths in as-synthesized PtxRu100−x NPs appear practically
indistinguishable from those in bulk Pt (2.77 Å) and bulk
Ru (2.67 Å), respectively. By contrast, metal-to-metal bond
lengths in post-synthesis treated NPs largely appear shorter
than those in as-synthesized NPs and respective solids indicating increased interactions between Pt and Ru metallic species.
Moreover, the post-synthesis treatment induced shortening of
metal-to-metal bond lengths appears irregular in a sense that
(i) it does not evolve linearly with NP bimetallic composition
and (ii) is more pronounced for atoms at NP surface. The postsynthesis treatment is also seen to change the way Pt and Ru
species arrange with respect to each other as manifested by
the significant diﬀerences between Pt–Pt, Pt–Ru and Ru–Ru
first CNs in as-synthesized (Table S7†) and respective fully activated Pt–Ru NPs (Table S8†). The change is particularly prominent with Pt31Ru69 NPs wherein a large number of surface Ru
atoms seem to have diﬀused into NP interior (surface Ru–Ru
CN drops from 5.7 to 5.2) and a corresponding number of subsurface Pt atoms moved toward NP surface (surface Pt–Pt CN
increase from 1.9 to 4.2) resulting in a sharpening of the
onion-like chemical pattern of the NPs, as exemplified in
Fig. 3 (second row, left) and S8.† Other studies48,49 on Pt–Ru
alloy NPs of low Pt content have also indicated a tendency of
Pt species to segregate at NP surface and explained it with the
fact that the surface energy of Pt (2.7 J m−2) is lower than that
of Ru (3.4 J m−2). The fact that not all but only a fraction of Pt
atoms segregates at NP surface and the rest cluster at NP
center may be considered a manifestation of the so-called
NP core–shell inversion eﬀect. As shown in prior work63,64
diﬀerent metallic species in binary alloy NPs may swap positions when the degree of atomic positional disorder and
strain, due to metal bond lengths mismatch, reach a critical
value. Often, a balance between the tendency of one of the two
types of metallic species (i.e. Pt) in binary alloy NPs to segregate at NP surface and the NP (Ru)core-(Pt)shell inversion
eﬀect is reached by formation of an “onion-like” chemical
pattern of the type shown in Fig. 3 (second row, left).65
First neighbor Pt–Pt and Ru–Ru distances, i.e. metal-tometal bond lengths, and first Pt–Pt and Pt–Ru CNs for Pt and
Ru atoms at the surface of fully activated for catalytic applications PtxRu100−x alloy NPs (x = 31, 49 and 75), as derived
from their 3D structure models of Fig. 3 (second row), are
shown in Fig. 5. Experimental data for the catalytic activity of
same NPs for gas-phase oxidation of CO and electro-oxidation
of ethanol, obtained as described in the Experimental section,
are also shown in the figure. Inspection of so-summarized
NP atomic structure vs. NP property data indicates that the catalytic activity of Pt–Ru alloy NPs studied here peaks when
Nanoscale, 2015, 7, 8122–8134 | 8129
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Paper
Fig. 5 (a) Catalytic activity of post-synthesis treated PtxRu100−x alloy
NPs (x = 31, 49 and 75) for gas-phase oxidation of CO given in terms of
the temperature T50 at which 50% conversion of CO is achieved. Note
the lower T50 the better the catalytic activity; (b) catalytic activity of
post-synthesis treated PtxRu100−x alloy NPs (x = 31, 49 and 75) for
electro-oxidation of ethanol given in terms of peak current generated as
a result of the reaction; (c) Pt–Pt and Ru–Ru bond lengths at the surface
of post-synthesis treated PtxRu100−x alloy NPs (x = 31, 49 and 75) and (d)
ﬁrst neighbor Pt–Pt and Pt–Ru CNs at the surface of same NPs. Note
data in (a) and (b) are obtained as described in the Experimental section.
Data in (c) and (d) are derived from RMC reﬁned structure models shown
in Fig. 3 (second row).
(i) Pt–Pt and Ru–Ru bond lengths at NP surface are, respectively, shortened and elongated most and (ii) surface Pt–Pt and
Pt–Ru first CNs are nearly identical. To comprehend the
observed NP atomic structure-catalytic activity relationships we
reasoned them within the framework of three most frequently
evoked explanations for the higher catalytic activity of metallic
alloy NPs as compared to that of respective monometallic NPs.
Those explanations argue that (i) diﬀerent metallic species at
the surface of alloy NPs do not interact but act cooperatively as
to promote particular steps of chemical reactions, (ii) diﬀerent
metallic species at the surface of alloy NPs interact by exchanging charge and/or adjusting size both aﬀecting their electronic structure in a way favorable for chemical reactions and
(iii) diﬀerent metallic species at NP surface assemble in particular configurations providing active catalytic sites.50–52 In
the case of Pt–Ru alloy NPs explanation (i) has been put
forward in terms of the so-called bi-functional mechanism53,54
envisioning formation of surface Ru-hydrate species easing CO
oxidation on nearby surface Pt atoms. The mechanism,
however, has been found incapable of explaining the improved
catalytic activity of Pt–Ru NPs wherein Pt atoms completely
cover NP surface thus screening all Ru atoms from chemical
species adsorbed at that surface.17,20,21 On the other hand,
8130 | Nanoscale, 2015, 7, 8122–8134
Nanoscale
explanation (ii) has found strong support in numerous experimental and theoretical studies on Pt–Ru alloy NPs indicating
the presence of charge transfer between Pt and Ru species55–59
accompanied by a considerable shortening of Pt–Pt and
elongation of Ru–Ru metal bond lengths.50,52,60–62 According
Pauling’s theory of chemical bonds66,67 some charge transfer
between Pt and Ru species may be expected due to their somewhat diﬀerent electronegativity that is 2.3 and 2.2, respectively.
Change in Pt–Pt and Ru–Ru bond lengths may also be
expected since in order for the atomic-level stresses in Pt–Ru
alloy NPs to be optimized the larger in size Pt species (2.77 Å)
should “shrink” and the smaller in size Ru species (2.67 Å) –
“expand” as that the ratio of their size/radii becomes as close
to one as possible.67,68 Changes in metal-to-metal bond
length, strength and the electronic structure of respective
metal species are known to be largely interrelated66,67 and so
diﬃcult to take apart. Nevertheless, it has been shown that
metal-to-metal bond length changes of the type shown in
Fig. 5c would induce a strong downshift of the d electron
states of surface Pt atoms in PtxRu100−x alloy NPs at x ∼ 50.
This would modify the energies of bonding of chemical
species to surface Pt atoms and so the activation barriers of
critical steps of chemical reactions, such as gas-phase oxidation of CO and electro-oxidation of ethanol, in a way
enhancing the catalytic activity of PtxRu100−x alloy NPs for
x ∼ 50.50,52,60–62 Indeed this is exactly what data in Fig. 5(a and b)
show. Explanation (iii) has been contemplated in several
studies on Pt–Ru alloy NPs26,50,58,69,70 but not recognized explicitly so far. Data in Fig. 5 provides strong evidence that not
only electronic (ii) but also atomic ensemble (iii) eﬀects, signified by the prevalence of atomic configurations where both Pt
and Ru metal species have pretty much the same number of
unlike first neighbors (Pt–Pt and Pt–Ru first CNs of 3.8 and
3.5, and Ru–Pt and Ru–Ru CNs of 3.5 and 3.8, respectively, see
Table S8†), contribute to the enhanced catalytic activity of
PtxRu100−x alloy NPs for x ∼ 50. The fact (see Fig. S8†) that
near surface atomic layers in PtxRu100−x alloy NPs are largely
stacked in a hcp-like manner for x ∼ 50 while those in PtxRu100−x
alloy NPs of lower (x = 31) and higher (x = 75) Pt content
appear fcc-like stacked is another indication that atomic
ensemble eﬀects, this time signified by distinct stacking of
atomic layers near NP surface, contribute to the enhancement
of catalytic activity of PtxRu100−x alloy NPs for x ∼ 50. Note,
electronic and atomic ensemble eﬀects described above do not
contradict but complement each other since it is when both
metallic species at the surface of binary alloy NPs are closely
packed altogether and have pretty much the same number of
unlike first neighbors to interact with the likelihood of (i)
charge transfer between the species and/or (ii) sizable adjusting of species’ size is maximized. Also, note, our findings do
not dismiss the “bi-functional” mechanism (i). Rather, they
highlight that it is the optimal combination of downshift of d
electron states of surface Pt atoms, surface Pt–Pt and Ru–Ru
metal bond lengths, surface Pt–Pt and Pt–Ru CNs and stacking
of near surface atomic layers in Pt100−xRux alloy NPs with
x ∼ 50 what enables an eﬃcient cross-reactivity of reaction
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
intermediates on the diﬀerent catalytic sites, such as transfer
of activated or intermediate CO-species to sites favoring a reaction with activated O-species, and/or refreshing the catalytic
sites, resulting in reducing the propensity of positioning.
With the relationships between atomic structure and catalytic activity of Pt–Ru alloy NPs established and understood well
useful clues to improving further the latter by tuning up the
former can be contemplated. For example, it may be conjectured that a synthesis and post-synthesis protocol aimed at (i)
stacking atomic layers near NP surface in a hcp-type manner,
(ii) densely populating those layers with Pt and Ru atoms so
that each of them would have a close to the maximum possible
number of unlike first neighbours and (iii) strengthening the
interactions between Pt and Ru species near NP surface as that
Pt–Pt and Ru–Ru metal bond lengths “contract” and “expand”,
respectively, may be a rather successful route to producing
Pt–Ru alloy NPs with improved catalytic activity for gas-phase
oxidation of CO and electro-oxidation of ethanol. Achieving
each of (i), (ii) and (iii) is important since the catalytic activity,
configuring of surface metallic species, surface metal-to-metal
bond lengths and strength, i.e. electronic structure of surface
metallic species, are intimately coupled.50,52,62,71–74 Also, it
may be conjectured that such a route may lead to catalysts
with improved durability since studies have shown that well
alloyed Pt–Ru NPs are more stable in catalytic reactions than
partially segregated ones.75,76 Furthermore, the route would
facilitate adding a third metallic species, such as Mo, W or Sn,
to the mix of Ru and Pt species as to stabilize certain quasiternary nanoalloys with excellent catalytic properties for
electro-oxidation of ethanol.77 Besides, the route would be
beneficial even if the “bi-functional mechanism” discussed
above were to be factored in. The route may follow the synthesis protocol employed here since it directly yields Pt–Ru
NPs of fcc/hcp metallic alloy type character. It may aim at Pt–
Ru NPs with Pt : Ru ratio close to 50 : 50 or somewhat lower if
the amount of scarce Pt in nanoalloy catalysts is to be reduced.
In the latter case, the post-synthesis treatment of Pt–Ru NPs
may need to be conducted so that Pt species segregate at NP
surface in a way consistent with goals (i), (ii) and (iii) formulated above. Indeed adjusting parameters of a post-synthesis
treatment of the type carried out here, such as duration, temperature and gas atmosphere of the treatment, is emerging as
a rather fruitful procedure for optimizing the atomic structure
of binary and ternary metal alloy NPs with respect to their catalytic properties.78,79
Paper
arrangement, size, and electronic structure of atoms at NP
surface, however, strongly depend on the way atoms in the
layer(s) beneath it are positioned. Therefore, detailed knowledge of the atomic-scale structure across metallic alloy NPs is
a prerequisite to achieving a better understanding of their catalytic properties and, hence, enabling a rational design
approach to producing and optimizing metallic alloy NPs for
catalytic applications.
Here we show how such knowledge can be obtained by total
scattering experiments coupled to atomic PDFs analysis and
3D computer modeling. Using Pt–Ru alloy NPs as an example,
we also show how, once obtained, it can be used for establishing atomic structure–catalytic properties relationships in metallic NPs. In particular, we show that electronic eﬀects,
resulting from increased interactions between Pt and Ru
species near NP surface, and practically unrecognized so far
atomic ensemble eﬀects, resulting from particular configuring
both of atomic layers and individual metallic species near NP
surface, contribute to an increase in the catalytic activity of
PtxRu100−x alloy NPs for gas-phase oxidation of CO and electrooxidation of ethanol at x ∼ 50. Furthermore, keep using Pt–Ru
alloy NPs as an example, we demonstrate how NP atomic structure-properties relationships can help steer synthesis and postsynthesis treatment of metallic NPs toward further improvement of their functionality. In particular, we suggest routes to
improving the catalytic activity of Pt–Ru alloy NPs by enhancing the electronic and atomic ensemble eﬀects in the NPs.
Based on results of present and prior studies16,39,46,47,80,81
we argue that the demonstrated here approach to establishing
and making use of synthesis-atomic structure–properties
relationships in metallic NPs is both widely applicable in
terms of metallic NP composition and environment and easily
accessible thanks to widespread instrumentation for total scattering experiments and already available computer power and
relevant software. It may be added that detailed knowledge of
the atomic structure across metallic NPs is very much needed
when NPs are explored not only for catalytic but also magnetic
and biomedical applications. For example, it can help estimate
the contribution of surface atoms to the so-called surface magnetism aﬀects82 and achieve a better understanding of how
proteins dock to the surface of metallic NPs,83 respectively.
Therefore, it may be envisaged that the approach to obtaining
such knowledge demonstrated here will find widespread utility
in many areas of science and technology of metallic NPs.
Acknowledgements
Conclusions
Catalytic activity of metallic NPs is governed by a series of
destruction and creation of bonds between NP surface atoms
and chemical species adsorbed at NP surface. Complete or
partial alloying of metal species at the nanoscale can modify
the electronic structure, hence, size (or vice versa) and arrangement of individual metallic species at NP surface thus influencing the catalytic activity of metallic NPs significantly. The
This journal is © The Royal Society of Chemistry 2015
Work on this paper was supported by DOE-BES Grant
DE-SC0006877. Work at the Advanced Photon Source was supported by DOE under Contract DEAC02-06CH11357.
References
1 B. R. Cuenya, Synthesis and catalytic properties of metal
nanoparticles: Size, shape, support, composition, and oxidation state eﬀects, Thin Solid Films, 2010, 518, 3127.
Nanoscale, 2015, 7, 8122–8134 | 8131
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Paper
2 X. Liu, D. Wang and Y. Li, Synthesis and catalytic properties of bimetallic nanomaterials with various architectures, Nano Today, 2012, 7, 448.
3 J. K. Norskov and C. H. Christensen, Toward eﬃcient
hydrogen production at surfaces, Science, 2006, 312, 1322.
4 S. Link, Z. L. Wang and M. A. El-Sayed, Alloy Formation of
Gold-Silver nanoparticles and the Dependence of the
Plasmon Absorption on their composition, J. Phys. Chem.,
1999, 103, 3529.
5 S. Link and M. A. El-Sayed, Optical Properties and ultrafast
dynamics of metallic nanocrystals, Annu. Rev. Phys. Chem.,
2003, 54, 331.
6 D. Antoniak, M. Spasova, A. Trunova, K. Fauth,
A. R. Wilhelm, J. Minar, H. Ebert, M. Farle and H. Wende,
Inhomogenous alloying in FePt nanoparticles as a reason
for reduced magnetic moments, J. Phys.: Condens. Matter,
2009, 21, 336002.
7 K. Kadau, M. Gruner, P. Entel and M. Kreth, Modeling
structural and magnetic phase transitions in Fe-Ni nanoparticles. Phase Transitions, Phase Transitions, 2003, 76,
355.
8 D. Alloyeau, C. Ricolleau, C. Mottet, T. Oitkawa,
C. Langlois, L. Y. Bouar, N. Braidy and A. Loiseau, Size and
shape eﬀects on the order-disorder phase transition in
CoPt nanoparticles, Nat. Mater., 2009, 8, 940.
9 P. Tartaj, M. d. P. Morales, S. Veintemillas-Verdaguer,
T. Gonzalez-Carreno and C. J. Serna, The preparation of
magnetic nanoparticles for applications in biomedicine,
J. Phys. D: Appl. Phys., 2003, 36, R182.
10 J. R. Lakowicz, Radiative decay engineering: Biophysical
and biomedical applications, Anal. Biochem., 2001, 298, 1.
11 C.-A. J. Lin, C.-H. Lee, J.-T. Hsiech, H. H. Wang, J. K. Li,
J. L. Shen, W. H. Chan, H.-I. Yeh and W. H. Chang,
Syntheis of fluorescent metallic toward biomedical application: recent progress and present challenges, Jpn. J. Med.
Electron. Biol. Eng., 2009, 29, 276.
12 J. M. Thomas and W.-J. Thomas, Principle and practice of
heterogeneous catalysis, VCN, 1997.
13 W. J. Huang, R. Sun, J. Tao, L. D. Menard, R. G. Nuzzo and
J. M. Zuo, Coordination-dependent surface atomic contraction in nanocrystals revealed by coherent diﬀraction, Nat.
Mater., 2008, 7, 308.
14 D. Marks, Experimental studies of small particle structures,
Rep. Prog. Phys., 1994, 57, 603.
15 A. Frenkel, Solving the 3D structure of nanoparticles,
Z. Kristallogr., 2007, 222, 605.
16 V. Petkov, B. Prasai, Y. Ren, S. Shan, J. Luo, P. Joseph and
Ch. J. Zhong, Solving the nanostructure problem: exemplified on metallic alloy nanoparticles, Nanoscale, 2014, 6,
10048.
17 L. Zhang, J. Kim, H. M. Cheng, F. Nah, K. Dudeck,
R.-S. Liu, G. A. Botton and J. Zhang, A novel CO-tolerant
PtRu core–shell structured electrocatalyst with rich Ru rich
in core and Pt rich in shell for hydrogen oxidation reaction
and its implication in PEMFCs, J. Power Sources, 2011, 196,
9117.
8132 | Nanoscale, 2015, 7, 8122–8134
Nanoscale
18 C.-Yu. Chen, W.-H. Lai, Ch.-Ch. Chen and S.-W. Hsu,
Eﬀects on nitrogen and CO concentrations on the performance of PEMFCs with Pt-Ru anode catalyst, J. Power Sources,
2013, 243, 138.
19 H. Kageyama, T. Ioroi, T. Kojima, H. Senoh, N. Takeichi,
K. Nomura and K. Tanimoto, XAFS Analysis of Pt and Pt-Ru
Catalysts for PEFCs by In situ Measurements under Operating Conditions in the Fluorescence Mode, AIP Conf. Proc.,
2007, 882, 645.
20 S. Alayoglu, A. U. Nilekar, M. Mavrikakis and B. Eichhorn,
Ru–Pt core–shell nanoparticles for preferential oxidation of
carbon monoxide in hydrogen, Nat. Mater., 2008, 7(4), 333.
21 Yu.-C. Hsieh, Yu. Zhang, D. Su, V. Volkov, R. Si, L. Wu,
Y. Zhu, W. An, P. Liu, P. He, S. Ye, R. R. Adzic and
J. X. Wang, Ordered bilayer ruthenium-platinum core–shell
NPs as CO-tolerant fuel cell catalysts, Nat. Commun., 2013,
4, 2466.
22 S.-Y. Huang, S.-M. Chang, Ch.-L. Lin, Ch.-H. Chen and
Ch.-T. Yeh, Promotion of the electrochemical activity of
a bimetallic Pt-Ru catalyst by oxygen-induced segregation,
J. Phys. Chem. B, 2006, 110, 23300.
23 S. Stoupin, E.-H. Chung, S. Chattopadhyay, C. U. Segre and
E. S. Smotkin, Pt and Ru X-ray absorption Spectroscopy of
PtRu anode catalysts in operating direct methanol fuel
cells, J. Phys. Chem. B, 2006, 110, 9932.
24 R. J. K. Witshire, C. R. King, A. Rose, P. P. Wells, H. Davies,
M. P. Hogarth, D. Thompsett, B. Theobald,
F. W. Mosselmans, M. Roberts and A. E. Russel, Eﬀects of
composition on structure and activity of PtRu/C catalysts,
Phys. Chem. Chem. Phys., 2009, 11, 2305.
25 E. Antolini, The problem of Ru dissolution from Pt-Ru catalysts during fuel cell operation: analysis and solutions,
J. Solid State Electrochem., 2011, 15, 455.
26 E. Antollini, Catalysts for direct ethanol fuel cells, J. Power
Sources, 2007, 170, 1; C. Lamy, A. Lima, V. Leruhun,
F. delime, Ch. Coutanceau and J.-M. Leger, Recent
Advances in the development of direct alcohol fuel cells,
J. Power Sources, 2002, 105, 283.
27 F. H. B. Lima and E. R. Gongalez, Ethanol electro-oxidation
on carbon supported Pt-Ru, Pt-Rh and Pt-Rh-Ru nanoparticles, Electrochim. Acta, 2008, 53, 2963.
28 T. Egami and S. J. L. Billinge, Underneath the Bragg peaks:
Structural Analysis of Complex Materials, Pergamon Press,
Elsevier Ltd, New York, 2003.
29 V. Petkov, Nanostructure by high-energy XRD, Mater. Today,
2008, 11, 28.
30 K. Page, T. C. Hood, T. Proﬀen and R. B. Neder, Building
and refining complete nanoparticle structures with total
scattering data, J. Appl. Crystallogr., 2011, 44, 327.
31 A. P. Sutton and J. Chen, Long-range Finnis-Sinclair potentials, Philos. Mag. Lett., 1990, 61, 139.
32 Y. Kimura, Y. Qi, T. Cagin and W. Goddard III, The
Quantum Sutton-Chen Many-body Potential for Properties
of fcc Metals, MRS Symp. Ser., 1999, 554, 43.
33 J. Yin, S. Shan, M. S. Ng, L. F. Yang, D. Mott, W. Fang,
N. Kang, J. Luo and C. J. Zhong, Langmuir, 2013, 29, 9249.
This journal is © The Royal Society of Chemistry 2015
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Nanoscale
34 J. Luo, N. Kariuki, L. Han, L. Wang, C. J. Zhong and T. He,
Preparation and characterization of carbon-supported
PtVFe electrocatalysts, Electrochim. Acta, 2006, 51, 4821.
35 P. Dash, T. Bond, C. Fowler, W. Hou, N. Coombs and
R. W. J. Scott, Rational Design of Supported PdAu
Nanoparticle Catalysts from Structured Nanoparticle Precursors, J. Phys. Chem. C, 2009, 113(29), 12719.
36 V. Petkov, RAD, a program for analysis of X-ray diﬀraction
data from amorphous materials for personal computers,
J. Appl. Crystallogr., 1989, 22, 387.
37 J. Yin, S. Shan, M. S. Ng, L. F. Yang, D. Mott, W. Fang,
N. Kang, J. Luo and C. J. Zhong, Catalytic and Electrocatalytic Oxidation of Ethanol over Palladium-Based Nanoalloy
Catalysts, Langmuir, 2013, 29, 9249.
38 J. M. Hutchinson, Platinum-Ruthenium Binary Alloy Phase
Diagram, Platinum Met. Rev., 1972, 16, 88.
39 V. Petkov, N. Bedford, M. R. Knecht, M. G. Weir,
R. M. Crooks, W. Tang, G. Henkelman and A. Frenkel,
Periodicity and Atomic Ordering in Nanosized Particles of
Crystals, J. Phys. Chem. C, 2008, 112, 8907.
40 V. Petkov, Y. Lee, S. Sun and Y. Ren, Noncrystallographic
Atomic Arrangement Driven Enhancement of the Catalytic
Activity of Au Nanoparticles, J. Phys. Chem. C, 2012, 116,
26668.
41 V. Petkov, R. Loukrakpam, L. Yang, B. N. Wanjala, J. Luo,
C. J. Zhong and S. Shastri, Pt-Au alloying at the nanoscale,
Nano Lett., 2012, 12, 4289.
42 L. Farrow, P. Juhas, J. Liu, D. Bryndin, E. Bozin, J. Bloch,
T. Proﬀen and S. Billinge, PDFfit2 and PDFgui: computer
programs for studying nanostructure in crystals, J. Phys.:
Condens. Matter, 2007, 19, 335219.
43 L. Deng, W. Hu, H. Deng and S. Xiao, Surface Segregation
and Structural Features of Bimetallic Au–Pt Nanoparticles,
J. Phys. Chem. C, 2010, 114(25), 11026.
44 H. Rafii-Tabar and A. S. Sutton, Long-range Finnis-Sinclair
potentials for fcc metallic alloys, Philos. Mag. Lett., 1991,
63, 217.
45 M. W. Finnis and J. E. Sinclair, A simple empirical N-body
potential for transition metals, Philos. Mag. A, 1984,
50, 45.
46 V. Petkov, S. Shan, P. Chupas, J. Yin, L. Yang, J. Luo and
C. J. Zhong, Noble-transition metal nanoparticle breathing
in a reactive gas atmosphere, Nanoscale, 2013, 5, 7379.
47 V. Petkov, S. Shastri, S. Shan, P. Joseph, J. Luo, C. J. Zhong,
T. Nakamura, Y. Herbani and S. Sato, Resolving Atomic
Ordering Diﬀerences in Group 11 Nanosized Metals and
Binary Alloy Catalysts by Resonant High-Energy X-ray Diﬀraction and Computer Simulations, J. Phys. Chem. C, 2013,
117, 22131.
48 A. V. Ruban, H. L. Skriver and J. K. Norskov, Surface segregation energies in transition-metal alloys, Phys. Rev. B:
Condens. Matter, 1999, 59, 15990.
49 Ch. W. Hill, M. S. Nasher, A. I. Frenkel, J. R. Shapley and
R. G. Nuzzo, Carbon support eﬀect on bimetallic Pt-Ru NPs
formed from molecular precursors, Langmuir, 1999, 15, 690
and refs. therein.
This journal is © The Royal Society of Chemistry 2015
Paper
50 B. Hammer and J. K. Norskov, Theoretical Surface Science
and catalysis – calculations and concepts, Adv. Catal., 2007,
45, 71.
51 F. Marooun, F. Ozanam, O. M. Magnusen and R. J. Behm,
The role of atomic assemblies in the reactivity of bimetallic
electrocatalysts, Science, 2001, 293, 1811.
52 J. K. Norskov, T. Bigaard, J. Rossmeisl and
C. H. Christiansen, Towards computational design of solid
catalysts, Nat. Chem., 2009, 1, 3746.
53 M. Watanabe and S. Motoo, Electrocatalysis by ad-atoms:
Part III. Enhancement of the oxidation of carbon monoxide
on platinum by ruthenium ad-atoms, J. Electroanal. Chem.
Interfacial Electrochem., 1975, 60, 267.
54 H. A. Gaisteiger, N. Markovic, P. N. Ross and E. J. Cairns,
CO Electrooxidation on Well-Characterized Pt-Ru Alloys,
J. Phys. Chem., 1994, 98, 617.
55 J. McBreen and S. Mukerjee, In situ X-ray absorption
studies of a Pt-Ru electrocatalyst, J. Electrochem. Soc., 1995,
142, 3399.
56 R. J. Wiltshire, C. R. King, A. Rose, P. P. Wells, H. Davies,
M.
P.
Hogarth,
D.
Thomsett,
B.
Theobald,
F. W. Mosselmans, M. Roberts and A. E. Russel, Eﬀects of
composition on structure and activity of PtRu/CC catalysts,
Phys. Chem. Chem. Phys., 2009, 11, 2305.
57 P. Hu, D. King, M. H. Lee and M. C. Payne, Orbital mixing
in CO chemisorption on transition metal surfaces, Chem.
Phys. Lett., 1995, 246, 73.
58 Q. Ge, S. Desai, M. Neurock and K. Koourtakis, CO adsoption on the surface of Pt-Ru bulk alloy, J. Phys. Chem., 2001,
105, 9533.
59 K.-W. Park, J-h. Choi, N.-K. Kwon, S.-A. Lee, Y.-E. Sung,
H.-Y. Ha, S.-A. Hong, H. Kim and A. Wieckowski, Chemical
and electronic eﬀects of Ni and Pt/Ru/Ni alloy particles in
methanol electrooxidation, J. Phys. Chem. B, 2002, 106,
1869.
60 A. Schlapka, M. Liska, A. Gros, U. Kasberger and P. Jacob,
Surface strain vs Substrate Interaction in Heteroepitaxial
Metal Layers: Pt on Ru (0001), Phys. Rev. Lett., 2003, 91,
016101.
61 M. Mavrikakis, B. Hammer and J. K. Norskov, Eﬀect of
strain on the reactivity of metal surfaces, Phys. Rev. Lett.,
1998, 81, 2819.
62 T. Bligaard and J. K. Nørskov, Ligand eﬀects in heterogeneous catalysis and electrochemistry, Electrochim. Acta,
2007, 52, 5512.
63 R. M. Anderson, L. Zhang, J. A. Loussaert, A. I. Frenkel,
G. Henkelman and R. M. Crooks, An experimental and
theoretical investigation of the inversion of Pd@Pt core@shell dendrimer encapsulated NPs, ACS Nano, 2013, 7,
9345.
64 S. G. Kwon, G. Krylova, P. J. Phillips, R. F. Klie,
S. Chattopadhyay, T. Shibata, E. E. Bunel, Y. Liu,
V. B. Prakapenka, B. Lee and E. V. Shevchenko, Heterogeneous nucleation and shape transformation of multicomponent metallic nanostructures, Nat. Mater., 2015, 14,
215.
Nanoscale, 2015, 7, 8122–8134 | 8133
View Article Online
Published on 10 April 2015. Downloaded by Central Michigan University on 24/04/2015 12:34:49.
Paper
65 A. V. Evteev, E. V. Levchenko, I. V. Belova and G. E. Murch,
Formation of a surface-sandwich structure in Pd-Ni NPs by
interdiﬀusion: atomistic modeling, Diﬀus. Fundam., 2007,
6, 81.
66 L. Pauling, The Nature of the Chemical Bond, Cornell University Press, Ithaca, NY, 1975.
67 L. Pauling, Factors Determining the Average Atomic
Volumes in Intermetallic Compounds, Proc. Natl. Acad.
Sci. U. S. A., 1987, 84, 4754.
68 T. Rajasekharam and V. Seshubai, Charge Transfer on the
Metallic Atom-Pair Bond, and the Crystal Structures
Adopted by Intermetallic Compounds, Acta Crystallogr.,
Sect. A: Fundam. Crystallogr., 2012, 68, 156.
69 L. Giorgi, A. Pozio, C. Bracchhini, R. Giori and S. Turtu, H2
and H2/CO activation mechanism on Pt/C, Ru/C and Pt-Ru/
C electrocatalysts, J. Appl. Electrochem., 2001, 31, 325.
70 S. Stoupin, E. Chung, S. Chattopadhyay, C. U. Segre and
S. S. Smotkin, Pt and Ru EXFS spectroscopy on PtRu anode
catalysts in operating direct methanol fuel cell, J. Phys.
Chem., 2006, 110, 9932.
71 A. Aduayo, G. Murrieta and R. de Coss, Elastic stability and
electronic structure of fcc Ti, Zr, and Hf: a first-principles
study, Phys. Rev. B: Condens. Matter, 2002, 65, 092106.
72 M. Wasniowska, W. Wulfhekel, M. Przybylski and
J. Kirschner, Submonolayer regime of Co epitaxy on
Pd(111): morphology and electronic structure, Phys. Rev. B:
Condens. Matter, 2008, 78, 035405.
73 A. L. Vazquez de Parga, F. J. Garcia-Vidal and R. Miranda,
Detectingg electronic states at stacking faults in magnetic
thin films by tunneling microscopy, Phys. Rev. Lett., 2000,
85, 4365.
74 J. J. Mortensen, Y. Morikawa, B. Hammer and
J. K. Norskov, Density functional theory of N2 adsorption
and dissociation on a Ru (0001) surface, J. Catal., 1997,
169, 85.
8134 | Nanoscale, 2015, 7, 8122–8134
Nanoscale
75 M.-S. Hyun, S.-K. Kim, B. Lee, D. Peck, Y. Shul and D. Jung,
Eﬀect of NaBH4 concentration on the characteristics of
PtRu/C catalyst for the anode of DMFC prepared by the
impregnation method, Catal. Today, 2008, 132(1–4), 138.
76 C. Bock, B. MacDougall and Y. LePage, Dependence of
CH3OH oxidation activity for a wide range of PtRu alloys:
Detailed analysis and new views, J. Electrochem. Soc., 2004,
151, A1269.
77 E. Antolini, Eﬀect of the structural characteristics of binary
Pt-Ru and ternary Pt-Ru-M fuel cell catalysts on the activity
of electro-oxidation of ethanol in acid medium, ChemSusChem, 2013, 6, 966.
78 J. Yin, S. Shan, L. Yang, D. Mott, O. Malis, V. Petkov, F. Cai,
M. S. Ng, J. Luo, B. H. Chen, M. Engelhard and C. J. Zhong,
Gold-Copper Nanoparticles: Nanostructural Evolution and
Bi-functional Catalytic Sites, Chem. Mater., 2012, 24, 4662.
79 B. N. Wanjala, B. Fang, S. Shan, V. Petkov, P. Zhu,
R. Loukrakpam, Y. Chen, J. Luo, J. Yin, L. Yang, M. Shao
and C. J. Zhong, Design of ternary nanoalloy catalysts:
eﬀect of nanoscale alloying and structural perfection on
electrocatalytic enhancement, Chem. Mater., 2012, 24,
4283.
80 M. A. Newton, K. W. Chapman, D. Thompsett and
P. J. Chupas, Chasing Changing nanoparticles with time
resolved Pair Distribution Function method, J. Am. Chem.
Soc., 2012, 134, 5036.
81 S. M. Oxford, P. L. Lee, P. J. Chupas, K. W. Chapman,
M. C. Kung and H. H. Kung, Study of supported PdAu bimetallic nanoparticles using in situ X-ray tools, J. Phys.
Chem., 2010, 114, 17085.
82 K. Sato, Magnetic nanoparticles: When atoms move
around, Nat. Mater., 2009, 8, 924–925.
83 G. Brancolini, D. B. Kokh, L. Calzolai, R. C. Wade and
S. Corni, Docking of ubiquitin to gold nanoparticles, ACS
Nano, 2012, 6, 9863.
This journal is © The Royal Society of Chemistry 2015